AWS Agents for Humans Hackathon

I got to know about the AWS Agents for Humans Hackathon, on Devpost.

I decided to form a team with a friend, who had some prior experience working with AWS, the team was formed, we even decided on what project to make for the hackathon.

However, by the time we finalized a idea less than a week was left on the timeline.

It was possible to make a project, using AI and Agentic IDE in the remaining time. However, I decided against it, my aim was to learn about AWS during the course of this hackathon. So the idea was dropped.

However I still wanted to learn AWS and Strands SDK, so instead of rushing to make a project I decided to divide the project into a multi-week endeavor, each week having numerous tasks.
As I make progress on each week, I will make a post about it.

LUMEN - Ambient AI companion for blind and low-vision users.

A phone-camera-driven, fully voice-operated multi-agent system that doesn’t just describe what’s in front of the user once — it builds a persistent memory of their environment over repeated use, so over time it behaves less like “describe this photo” and more like a companion who actually knows their home, their regular walk to the bus stop, the layout of their office. It proactively interrupts only for real hazards (stairs, an open door, a step down) and otherwise stays quiet.

Week 0 - Repository & production structure
Before moving ahead with the actual project, a Github Repo will be created, this will follow a typical production structure - MIT Licence, README.md, CHANGELOG.md, branch protection on 'main'.

As the progress is completed for each week, updates will be posted. So if you are curious to see the progress I make, you can follow me.